HPC-GPT: Integrating Large Language Model for High-Performance Computing
Explore the source record for details and available documents.
SEARCH · Engineering Papers
Search indexed NASA NTRS and DOE OSTI research on propulsion, heat transfer, battery materials and energy systems. Follow report and document links to the original sources.
Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.
Explore the source record for details and available documents.
None
Production LLM serving must simultaneously deliver high throughput, low latency, and sufficient context capacity under non-stationary traffic and mixed request requirements. Data parallelism (DP) maximizes throughput by running independent replicas, while tensor parallelism (TP) reduces per-request latency and pools memory for long-context inference. However, existing serving stacks typically commit to a static parallelism configuration at deployment; adapting to bursts, priorities, or long-context requests is often disruptive and slow. We present Flying Serving, a vLLM-based system that enables online DP-TP switching without restarting engine workers. Flying Serving makes reconfiguration practical by virtualizing the state that would otherwise force data movement: (i) a zero-copy Model Weights Manager that exposes TP shard views on demand, (ii) a KV Cache Adaptor that preserves request KV state across DP/TP layouts, (iii) an eagerly initialized Communicator Pool to amortize collective setup, and (iv) a deadlock-free scheduler that coordinates safe transitions under execution skew. Across three popular LLMs and realistic serving scenarios, Flying Serving improves performance by up to 4.79 × under high load and 3.47 × under low load while supporting latency- and memory-driven requests.
Code allows users to easily pull publications from PubMed, OSTI, and SCOPUS; identify biological terms and extract their contexts from these papers; and determine which terms are related to other terms using one of the eleven wrapped algorithms
This is a processed molecular dataset from this https://doi.ccs.ornl.gov/ui/doi/348 adding up to 50M molecules for the training and 486K molecules for the validation. Instructions on how to use/run/train this dataset can be found here: https://code.ornl.gov/candle/mlmol
CoVID-19 (cord19) open data source for training knowledge graph with LLM
Abstract Not Provided
The project developed a domain specific translator enable legacy MPI source code to tolerate communication delays, which are increasing over time due to technological factors. The translator performs source-to-source translation that incorporates semantic information into the translation process. The output of the translator is a C program runs as a data driven program, and uses an existing run time to overlap communication automatically
This document provides a short overview of cybersecurity procurement and implementation guidelines for federal facilities to complement the Task Order Request for Proposal (TO RFP) used by an ordering agency to communicate agency-, site-, and project-specific terms and conditions.
QSCOUT is the Quantum Scientific Computing Open User Testbed, a trapped-ion quantum computer testbed realized at Sandia National Laboratories on behalf of the Department of Energy's Office of Science and its Advanced Scientific Computing Research (ASCR) program.
Abstract not provided.
Abstract not provided.
Abstract not provided.
Abstract not provided.
Abstract not provided.
Abstract not provided.
It is an invariant that critical cyberphysical systems should not fail. Mission assurance requires systems to behave with predictability, especially in their ability to satisfy real-time constraints. The NIST standard for Engineering Trustworthy Secure Systems, National Institute of Standards and Technology (NIST) Special Publication (SP) 800-160v1r1 states that formal methods are the highest level for meeting assurance requirements. There is a tension between the formal method software development process (Figure 1), which does not introduce time specificity until the latter stages of the development process (concrete model), and the need to gain confidence that time constraints will be satisfied. This paper formalizes the practical application of temporal entities e.g. Propositional Temporal Logic (PTL), Temporal Logic of Actions Plus (TLA+), etc. to critical systems.